library(dplyr)
library(leaflet)
library(lubridate)
library(plotly)
library(ggplot2)
df_redq_coord = read.csv('./data_source/redq_coord.csv')
leaflet(data = df_redq_coord) %>%
addProviderTiles("Esri.NatGeoWorldMap") %>%
addCircleMarkers(~longitude, ~latitude, radius = 1, color = "blue", fillOpacity = 0.3)
df_1 = read.csv('./data_source/facilities-v1.csv')
df_1 %>%
filter(city=='New York') -> df_1_ny
distinct_df = df_1_ny %>% count(facsubgrp)
Pontos de metro de Ny
df_2 = read.csv('./data_source/NYC_Transit_Subway_Entrance_And_Exit_Data.csv')
filter(df_redq_coord, point=='pp1') -> pp1
filter(df_redq_coord, point=='pp2') -> pp2
filter(df_redq_coord, point=='pp3') -> pp3
filter(df_redq_coord, point=='pp4') -> pp4
train %>%
filter(pickup_latitude>=df_redq_coord$latitude[[1]][[1]]) %>%
filter(pickup_latitude<=df_redq_coord$latitude[[2]][[1]]) %>%
filter(pickup_longitude>=df_redq_coord$longitude[[1]][[1]]) %>%
filter(pickup_longitude<=df_redq_coord$longitude[[3]][[1]])-> df_pickup_dentroquadr
train %>%
filter(dropoff_latitude>=df_redq_coord$latitude[[1]][[1]]) %>%
filter(dropoff_latitude<=df_redq_coord$latitude[[2]][[1]]) %>%
filter(dropoff_longitude>=df_redq_coord$longitude[[1]][[1]]) %>%
filter(dropoff_longitude<=df_redq_coord$longitude[[3]][[1]])-> df_dropoff_dentroquadr
df_1_ny %>%
filter(latitude>=df_redq_coord$latitude[[1]][[1]]) %>%
filter(latitude<=df_redq_coord$latitude[[2]][[1]]) %>%
filter(longitude>=df_redq_coord$longitude[[1]][[1]]) %>%
filter(longitude<=df_redq_coord$longitude[[3]][[1]])-> df_ny_dentroquadr
df_2 %>%
filter(Station.Latitude>=df_redq_coord$latitude[[1]][[1]]) %>%
filter(Station.Latitude<=df_redq_coord$latitude[[2]][[1]]) %>%
filter(Station.Longitude>=df_redq_coord$longitude[[1]][[1]]) %>%
filter(Station.Longitude<=df_redq_coord$longitude[[3]][[1]])-> df_nysubway_dentroquadr
Mapa das corridas que começaram no quadrante
set.seed(22)
# Criando uma sequencia de 1 ate a quantidade de total de linhas
linhas.idx <- seq_len(nrow(df_pickup_dentroquadr))
# Obtendo aleatoriamente 10000 amostras de linhas do dataset
linhas.sample <- sample(linhas.idx, 1000)
# Amostra aleatoria com 10000 dados
df_pickup_dentroquadr[linhas.sample, ] -> df_pickup_dentroquadr_sample
leaflet(data = df_pickup_dentroquadr_sample) %>%
addProviderTiles("Esri.NatGeoWorldMap") %>%
addCircleMarkers(~ pickup_longitude, ~pickup_latitude, radius = 1, color = "blue", fillOpacity = 0.3)
Mapa das corridas
O pontos azuis são o inicio da corrida (pickup), os pontos vermelhos são as chegadas (dropoff) e os verde são pontos de interesse.
df_ny_dentroquadr %>%
dplyr::select('city','facname','facgroup')
df_nysubway_dentroquadr %>%
dplyr::select('Station.Name','Station.Latitude','Station.Longitude')
As viagens que pegam os passageiros (pickoff, azul no grafico) tem uma concentração maior na entrada da Penn Station, pois provavelmente as pessoas estão pegando taxi vindo do transporte público.
# df_pickup_dentroquadr_sample %>%
# plot_ly(x=~trip_duration,type="histogram")
#
# df_dropoff_dentroquadr_sample %>%
# plot_ly(x=~trip_duration,type="histogram") %>%
# layout(barmode = "overlay",xaxis=list(range=c(0,5000)))
plot_ly(alpha = 0.5) %>%
add_histogram(x = df_pickup_dentroquadr_sample$trip_duration) %>%
add_histogram(x = df_dropoff_dentroquadr_sample$trip_duration) %>%
layout(barmode = "overlay",xaxis=list(range=c(0,5000)))
#sd(df_pickup_dentroquadr_sample$trip_duration)
#sd(df_dropoff_dentroquadr_sample$trip_duration)
p1 <- df_pickup_dentroquadr_sample %>%
group_by(passenger_count) %>%
count() %>%
ggplot(aes(passenger_count, n, fill = passenger_count)) +
geom_col() +
scale_y_sqrt() +
theme(legend.position = "none")
plot(p1)
p4 <- df_pickup_dentroquadr_sample %>%
mutate(wday = wday(pickup_datetime, label = TRUE)) %>%
group_by(wday, vendor_id) %>%
count() %>%
ggplot(aes(wday, n, colour = vendor_id)) +
geom_point(size = 4) +
labs(x = "Day of the week", y = "Total number of pickups") +
theme(legend.position = "none")
plot(p4)
p5 <- df_pickup_dentroquadr_sample %>%
mutate(hpick = hour(pickup_datetime)) %>%
group_by(hpick, vendor_id) %>%
count() %>%
ggplot(aes(hpick, n, color = vendor_id)) +
geom_point(size = 4) +
labs(x = "Hour of the day", y = "Total number of pickups") +
theme(legend.position = "none")
plot(p5)
Preparação dos dados
source('preprocessing.R')
## Adiciona distância Euclidiana calculada a partir das coordenadas (arquivo Preprocessing.R)
#Distancia em KM
df_pickup_dentroquadr_sample$dist_euclidiana = dist_eucl(df_pickup_dentroquadr_sample)
## Adiciona distância de Manhattan calculada a partir das coordenadas (arquivo Preprocessing.R)
df_pickup_dentroquadr_sample$dist_manhattan = dist_manh(df_pickup_dentroquadr_sample)
df_pickup_dentroquadr_sample$velocidade = df_pickup_dentroquadr_sample$dist_manhattan / df_pickup_dentroquadr_sample$trip_duration
## Prepara data e hora da paprtida
### Com isso é possÃvel pegar horário de pico e dia da semana
df_pickup_dentroquadr_sample$pickup_hour <- hour(df_pickup_dentroquadr_sample$pickup_datetime)
df_pickup_dentroquadr_sample$pickup_month <- month(df_pickup_dentroquadr_sample$pickup_datetime)
df_pickup_dentroquadr_sample$pickup_weekdays <- wday(df_pickup_dentroquadr_sample$pickup_datetime)
Regração linear
linearMod <- lm(dropoff_latitude ~ dropoff_longitude, data=df_dropoff_dentroquadr_sample)
print(linearMod)
##
## Call:
## lm(formula = dropoff_latitude ~ dropoff_longitude, data = df_dropoff_dentroquadr_sample)
##
## Coefficients:
## (Intercept) dropoff_longitude
## 42.89290 0.02896
linearMod <- lm(pickup_latitude ~ pickup_longitude, data=df_pickup_dentroquadr_sample)
print(linearMod)
##
## Call:
## lm(formula = pickup_latitude ~ pickup_longitude, data = df_pickup_dentroquadr_sample)
##
## Coefficients:
## (Intercept) pickup_longitude
## 20.2273 -0.2774